CauseNet: Towards a causality graph extracted from the web

The sample set contains:

    {
        "causal_relation": {
            "cause": {
                "concept": "boom"
            },
            "effect": {
                "concept": "bust"
            }
        }
    }

It's practically a hedge-fund-in-a-box.

Plus, regardless of what you might think of how valid that connection is, what they're actually collecting, absent any kind of mechanism, is a set of all apparent correlations...